A third-generation PC-based computer model of alcohol use and abuse is being developed at PRC. The model will reproduce the complex, 20-year history of San Diego County with regard to alcohol retail activity, alcohol consumption patterns, drinking & driving behavior, social norms, and regulatory controls. Published research findings, survey data, and results from secondary data analyses are used to define and mathematically specify relationships among variables within and across the model subsystems. Through an iterative process of design, congruence testing, and redesign, the model is being refined toward the future point when it can be offered to researchers and prevention specialists as a reliable, comprehensive tool for understanding complex community dynamics and for estimating impacts of interventions intended to reduce alcohol- related problems. Once congruence between model results and actual data has been achieved, the San Diego County model will be used to simulate the impact over the period 1992-2001 of alternative feasible prevention interventions. The current project (beginning December 1, 1992 and ending November 30, 1997) proposes to answer two important research questions: (1) will the model yield congruent results when applied to other communities? and (2) what structural changes are needed to insure the transferability of the model to other communities? A secondary question is: if congruency and transferability can be established, what additional model enhancements are required, if any, to encourage communities to use the model as a tool for evaluating alternative prevention options prior to actual program implementation? These questions will be addressed in a fourth-generation effort. The model will be tested and refined in 12 diverse communities. the communities will be selected according to three criteria: local population composition, state alcohol beverage control structure, and type of community organization focused on prevention programming. Each year, the model will first be initialized with local data from three communities, then benchmark-tested for congruence. Once the model performs with demonstrated accuracy for a target community, local prevention specialists will be trained in its use. Subsequent local use of the model will be tracked and documented as part of a process evaluation aimed at (a) determining under what conditions the model is most frequently and effectively used and (b) identifying ways to make the model more user-friendly and useful, without sacrificing its congruency and transferability.